J. Paediatr. Child Health (2005) 41, 479–483
Cerebral palsy in Victoria: Motor types, topography and gross
motor function
Jason Howard,1 Brendan Soo,1,4 H Kerr Graham,1,4,5 Roslyn N Boyd,1,2,4 Sue Reid,3,4 Anna Lanigan,3,4 Rory Wolfe6
and Dinah S Reddihough3,4,5
Royal Children’s Hospital, Departments of 1 Orthopaedics, Neonatology2 , and Child Development and Rehabilitation3 , Murdoch
Childrens Research Institute4 , Departments of Paediatrics, University of Melbourne5 , and Epidemiology and Preventitive Medicine,
Monash University6 , Melbourne, Victoria, Australia
Objectives: To study the relationships between motor type, topographical distribution and gross motor function in a large,
population-based cohort of children with cerebral palsy (CP), from the State of Victoria, and compare this cohort to similar
cohorts from other countries.
Methods: An inception cohort was generated from the Victorian Cerebral Palsy Register (VCPR) for the birth years 1990–
1992. Demographic information, motor types and topographical distribution were obtained from the register and supplemented
by grading gross motor function according to the Gross Motor Function Classification System (GMFCS).
Results: Complete data were obtained on 323 (86%) of 374 children in the cohort. Gross motor function varied from GMFCS
level I (35%) to GMFCS level V (18%) and was similar in distribution to a contemporaneous Swedish cohort. There was a
fairly even distribution across the topographical distributions of hemiplegia (35%), diplegia (28%) and quadriplegia (37%) with
a large majority of young people having the spastic motor type (86%).
Conclusions: The VCPR is ideal for population-based studies of gross motor function in children with CP. Gross motor function
is similar in populations of children with CP in developed countries but the comparison of motor types and topographical
distribution is difficult because of lack of consensus with classification systems. Use of the GMFCS provides a valid and
reproducible method for clinicians to describe gross motor function in children with CP using a universal language.
Key words: cerebral palsy; classification system; gross motor function; motor skills; motor type; population register; topographical distribution.
Cerebral palsy (CP) is the most common cause of physical disability affecting children in Australia and most developed countries, with a prevalence of approximately 2 per 1000 live births.1
It has been described as ‘an umbrella term covering a group
of non-progressive, but often changing, motor impairment syndromes secondary to lesions or anomalies of the brain arising
in the early stages of its development’.2 Although these clinical syndromes are often not pure, recognition of the dominant
motor types and topography has been important for research
into causal pathways and possible prevention, correlation with
brain imaging, and for establishing a prognosis and setting management goals and strategies. It may also trigger a search for
associated problems such as epilepsy, feeding difficulties and
silent hip subluxation, which may significantly impact on the
well-being and development of these children.3,4 Traditionally,
CP has been classified according to motor type, topographical
distribution and functional severity,1–3 but as yet there has been
no consensus reached on either the descriptors or the definitions
of motor type and topographical distribution.
The motor type is usually described as spastic, dyskinetic,
ataxic, hypotonic or mixed. Currently in the USA, under the
auspices of the National Institutes for Health, a taskforce on
childhood motor disorders is working on the important issue of
motor type classifications and some helpful guidelines have been
published.5–7 The Surveillance of CP in Europe has also gone
through a consensus process to develop standard definitions and
classifications of topography and motor type with the use of
a training CD-ROM. However, although classical presentations
are easily recognized, there are many children with mixed or
changing motor types that are difficult to define.1,8
Classifications according to topographical distribution are
widely employed. Although hemiplegia, diplegia and quadriplegia are commonly used terms, monoplegia and triplegia sometimes exist as separate entities or may be grouped with hemiplegia and quadriplegia, respectively.1 The expressions double
hemiplegia, spastic tetraplegia and total body involvement are
sometimes used to describe a child with four-limb involvement
and with the upper limbs more severely involved than the lower
limbs. Not surprisingly, classifications based on motor type and
topographical distribution have poor reliability, even when observers are experienced and undergo special training.1,8
The most useful development in the classification of CP in recent years has been the development of the Gross Motor Function
Classification System (GMFCS).9 The GMFCS is a five-level
ordinal grading system based on the assessment of self-initiated
movement with emphasis on function during sitting, standing
and walking. Distinctions between different levels are based on
functional limitations, the need for walking aids or wheeled mobility and quality of movement (Table 1). Unlike the classification of motor type and topography, the GMFCS has been shown
to be a valid, reliable, stable and clinically relevant method for
the classification and prediction of motor function in children
with CP between the ages of 2 and 12 years.9,10
The main aims of this study were to determine the distribution of motor impairment in a large population cohort of young
people with CP with respect to the motor type, topographic distribution and gross motor function according to the GMFCS,
and to explore the relationships between the three classification methods. The determination of the spectrum of functional
limitation that characterizes the group of children with CP is
Correspondence: Assoc Professor Dinah S Reddihough, Department of Child Development and Rehabilitation, Royal Children’s Hospital, Flemington
Road, Parkville, Vic. 3052, Australia. Fax: +61 3 9345 5871; email: [email protected]
Accepted for publication 21 March 2005.
480
J Howard et al.
Table 1 Gross Motor Function Classification System (GMFCS) levels, for children aged 6–12 years9
GMFCS level I. Children walk indoors and outdoors, and climb stairs without limitation. Children perform gross motor skills including running and
jumping but speed, balance and coordination are impaired.
GMFCS level II. Children walk indoors and outdoors, and climb stairs holding onto a railing but experience limitations walking on uneven surfaces
and inclines and walking in crowds or confined spaces. Children have at best only minimal ability to perform gross motor skills such as running and
jumping.
GMFCS level III. Children walk indoors or outdoors on a level surface with an assistive mobility device. Children may climb stairs holding onto a
railing. Children may propel a wheelchair manually or are transported when travelling for long distances or outdoors on uneven terrain.
GMFCS level IV. Children may continue to walk for short distances on a walker or rely more on wheeled mobility at home, school and in the
community. Children may achieve self-mobility using a power wheelchair.
GMFCS level V. Physical impairments restrict voluntary control of movement and the ability to maintain antigravity head and trunk postures. All
areas of motor function are limited. Children have no means of independent mobility and are transported.
essential to enable planning for resource allocation and to facilitate studies relating to aetiology, prevention or prognosis.1
A secondary aim of this study was to compare our study cohort with recently published cohorts that have included motor
type, topographical distribution and GMFCS levels. Gorter et al.
reported the relationships between limb distribution (topographical distribution) motor impairment and functional classification
in 657 children, aged from 1 to 13 years at study onset, from
Ontario, Canada.11 Although not based on a CP register, this
is a large community-based sample from the originators of the
GMFCS. The study by Nordmark et al. is smaller (n = 167), but
is population-based and examines gross motor function, type of
motor impairment and topography of children with CP within
the public health system of southern Sweden.12 The time frame
of this Swedish study (birth years 1990–1993) is the same as the
current study.
METHODS
The Victorian Cerebral Palsy Register (VCPR) was used to identify an inception cohort of children with a confirmed diagnosis
of CP and born between January 1990 and December 1992 in the
State of Victoria, Australia. These 3 years were chosen to give
a large population cohort of children from a period of high-case
ascertainment for the register and to allow sufficient followup time to determine the true extent of motor impairment and
musculoskeletal issues.1,3
The VCPR is administered in accordance with state and institutional legislation governing privacy and ethical standards,
and attempts to include all children with CP born in the State
of Victoria since 1970. Ascertainment of cases is based on a
standard definition of CP.2 Multiple sources of ascertainment
are used, including records of patients who access a variety of
services at the major paediatric hospitals in Victoria, paediatricians and parent or consumer groups. Motor type, topographical
distribution and severity are entered into the VCPR at the time of
case ascertainment and confirmed at age 5 years, as is the practice with several other established CP registers.1 The GMFCS
level has recently been added to the VCPR dataset, and is now
included in the routine assessment of children in developmental
and orthopaedic clinics and in the gait laboratory.
For this study, a cohort of children was selected from the
VCPR using the following inclusion criteria:
1. Birth date between January 1990 and December 1992,
inclusive;
2. Adequate clinical records to allow determination of functional level according to the GMFCS, between the ages of 6
and 12 years; and
3. Survival to age 6 years.
Children were grouped according to motor type, topographic
pattern and GMFCS level. Motor types were classified as spastic,
dyskinetic, mixed, ataxic and hypotonic, as defined in Table 2.5–7
Spastic patterns were further classified according to topographical distribution as hemiplegia (unilateral involvement), diplegia
(bilateral involvement, with the lower limbs more affected than
the upper limbs) and quadriplegia (bilateral involvement with
the upper limbs more or equally involved). Motor type and topographical distribution, as entered on the VCPR, were checked
with the child’s case records and any discrepancies were resolved by discussion with the child’s developmental paediatrician, neurologist or physiotherapist. The motor types were also
cross-checked with records from the gait laboratory, where available, and discrepancies were resolved by a full review of clinical
records and gait laboratory video tapes and data files.
Children were also classified according to the GMFCS for
ages 6 to 12 years. Although the majority of children had already been assigned a GMFCS level independent of this study,
multiple records were checked for agreement and discrepancies
were resolved by discussion with the child’s physiotherapist and
developmental paediatrician. Children without an assigned GMFCS level were classified by at least two of the study authors after discussion with the child’s community and/or hospital-based
physiotherapist.
Stata 8.0 (Stata 2003) was used for the statistical analysis. To
explore the relationships between motor type and GMFCS, and
topographic distribution and GMFCS, ordered logistic regression was used to calculate odds ratios (OR) and 95% confidence
intervals (95% CI) that summarize the propensity for one motor
Table 2 Definitions of motor types
Spasticity
Dyskinesia
Mixed motor types
Ataxia
Hypotonia
Hypertonia in which resistance to externally imposed movement increases with increasing speed and varies with direction of
movement and/or rises rapidly above a threshold speed.
Involuntary, sustained or intermittent muscle contractions causing twisting and repetitive movements, abnormal postures or
both.
Clinical features of more than one type, usually spastic and dyskinetic.
Abnormal pattern of posture and/or movement with loss of orderly muscle coordination so that movements are performed
with abnormal force, rhythm or accuracy.
Abnormally low tone, in the trunk and limbs that must be distinguished from weakness.
Cerebral palsy in Victoria
481
Table 3 The distribution of motor types and topographical distribution in 323 children with cerebral palsy, born in Victoria between 1990 and 1992.
For each motor type/topographic pattern, the frequency and percentage are shown within each GMFCS level
Motor type/topographic pattern
Hypertonia
Spastic
Hemiplegia
Diplegia
Quadriplegia
Dyskinesia
Mixed
Ataxia
Hypotonia
Total cohort
GMFCS
Total cohort
I
n (%)
II
n (%)
III
n (%)
IV
n (%)
V
n (%)
n (%)
79 (81)
26 (33)
1 (1)
0 (0)
1 (5)
4 (44)
3 (33)
114 (35)
17 (17)
26 (33)
3 (3)
2 (40)
3 (14)
2 (22)
0 (0)
53 (16)
1 (1)
22 (28)
13 (13)
1 (20)
5 (24)
3 (33)
1 (11)
46 (14)
0 (0)
4 (5)
36 (35)
2 (40)
7 (33)
0 (0)
3 (33)
52 (16)
1 (1)
0 (0)
50 (48)
0 (0)
5 (24)
0 (0)
2 (22)
58 (18)
98 (30)
78 (24)
103 (32)
5 (2)
21 (7)
9 (3)
9 (3)
323 (100)
type or topographical distribution to score higher up the GMFCS
scale than a reference group. Proportions of children in different
groups were compared between studies using chi-squared tests,
and shifts up the GMFCS scale were summarized using OR from
an ordered logistic regression model.
and was non-ambulant because of a mixture of severe cognitive,
behavioural and motor deficits. The children graded as GMFCS
levels I and II who had spastic quadriplegia were noted to have
RESULTS
There were 374 children on the register with a birth date between
January 1990 and December 1992. Nine of these children who
died before the age of 6 years were excluded from our analysis.
A further 26 children were excluded due to inadequate records
and 16 more were lost to follow-up. This left 323 children for
analysis, 86% of the cohort. The distributions of motor type,
topographic pattern and GMFCS level are shown in Table 3.
Eighty-six per cent of children had a spastic motor type, and
were relatively evenly distributed between spastic hemiplegia,
diplegia and quadriplegia. The other groups with hypertonia
(dyskinetic, mixed) accounted for 8% of cases while the remaining motor types (ataxic, hypotonic) represented 6% of the
cohort. With respect to the GMFCS classification, level I was
the most common at 35%. The remaining cases were distributed
fairly equally from levels II to V, in the range of 14–18% per
level.
Within the spastic group, differences in motor function among
the three topography groups were extremely clear-cut. Compared to children with a hemiplegic distribution, children with
diplegia were more severe on the GMFCS scale (OR = 9.3, 95%
CI: 4.7, 18.0). Children with spastic quadriplegia had the lowest levels of function, being significantly higher on the GMFCS
scale than those with hemiplegia (OR = 341, 95% CI: 246, 1800)
and diplegia (OR = 72, 95% CI: 30, 172) (Fig. 1). In comparison
with the total group with spasticity, children with both dyskinesia (OR = 1.6, 95% CI: 0.4, 6.2) and hypotonia (OR = 1.8, 95%
CI: 0.5, 6.0) appeared to have slightly higher levels of function on the GMFCS scale, but the numbers of children in these
groups were too small for a conclusive finding. Children classified as having a mixed type of motor disorder were significantly
more severe on the GMFCS scale than children with spasticity
(OR = 2.8, 95% CI: 1.3, 5.8) or ataxia (OR = 5.3, 95% CI: 1.4,
20.1).
The exception to these general relationships between topographical distribution and GMFCS was one child with spastic
hemiplegia who was graded as GMFCS level V and three children with spastic quadriplegia classified as GMFCS levels I and
II. The child with spastic hemiplegia who was graded as GMFCS level V had a hemispherectomy for intractable epilepsy,
Fig. 1 GMFCS level for 279 children with spastic motor type from
a cohort of cerebral palsy cases in Victoria born between 1990 and
1992. (A) GMFCS for children with spastic hemiplegia; (B) GMFCS
for children with spastic diplegia; (C) GMFCS for children with spastic
quadraplegia.
482
J Howard et al.
Table 4 Motor type, topographical distribution and GMFCS levels in children with cerebral palsy in Victoria, Australia; Ontario, Canada; and Southern
Sweden
GMFCS
Victoria
Ontario
Sweden
I
114
183
68
Motor type
Victoria
Ontario
Sweden
II
35.3%
27.9%
40.7%
16.4%
12.2%
18.6%
Spastic
279
500
NA
46
122
23
5
39
35.1%
15.3%
40.0%
52
137
19
Mixed
1.5%
6.1%
21
58
Hemiplegia
98
98
50
IV
14.2%
18.6%
13.8%
Dyskinetic
86.4%
78.2%
Topography
Victoria
Ontario
Sweden
53
80
31
III
V
16.1%
20.9%
11.4%
Ataxic
6.5%
9.1%
9
16
2.8%
2.5%
Diplegia
78
217
58
58
135
26
18.0%
20.5%
15.6%
Hypotonic
9
26
2.8%
4.1%
Quadriplegia
28.0%
33.9%
46.4%
103
325
17
Total
36.9%
50.8%
13.6%
323
657
167
Total
323
639
Total
279
640
125
NA, not applicable.
bilateral upper motor neuron signs in both the upper and lower
limbs, principally hyper-reflexia and upgoing plantar responses.
However, these three children had good selective motor control,
good strength and functioned at a very high level.
Comparison with two other cohorts
The GMFCS levels, motor types and topographical distributions
for the three cohorts are shown in Table 4. The Swedish study,
like ours, was a cohort derived from a population-based registry.
There was a shift in topography towards four limb involvement
in our cohort relative to the Swedish cohort (OR = 1.8, 95%
CI: 1.2, 2.7, P = 0.002), but the distributions of GMFCS were
similar (OR = 0.76, 95% CI: 0.54, 1.07). It was not possible to
compare motor types as the Swedish study followed a different
classification system.
In the Canadian study, a random sample of eligible children was selected that was large enough to ensure at least 60
children in each GMFCS level. There were significant differences in topography compared to both Sweden and Victoria
(χ 2 (d.f. = 4) = 94.2, P < 0.001); and across the motor types
(χ 2 (d.f. = 4) = 14.3, P = 0.007). In particular, the Victorian cohort had a higher percentage of spastic motor types (86%) than
the Ontario cohort (78%) and a lower percentage of dyskinesia
(1.5% vs 6.1%). Compared to the cohort from Ontario, the Victorian cohort (OR = 0.71, 95% CI: 0.56, 0.91) and the Swedish
cohort (OR = 0.55, 95% CI: 0.40, 0.74) were both relatively less
severe on the GMFCS.
DISCUSSION
The widespread adoption of the definition of CP proposed by
Bax in 1964 has resulted in the grouping together of large numbers of children with a wide variety of movement disorders, topographical distributions and functional abilities under the ‘cerebral palsy umbrella’.3,13 The marked variation in functional abilities of children sharing the common diagnosis of CP has led to
repeated efforts to classify children into meaningful clinical syndromes. In terms of motor type and topographical distribution,
these efforts have had limited success, and this makes it difficult
to make meaningful time and geographical comparisons.1–3,8,9
By contrast, the use of the GMFCS now facilitates a reliable
means of classification of the severity of the motor disorder. The
distribution of children throughout all GMFCS levels confirms
the wide range of function and disability in a typical population
sample of children with CP. GMFCS level I children are completely independent, they do not use aids, and they usually have
mild spastic hemiplegia. By contrast, children in GMFCS level
V have no independent mobility and often have severe spastic
dyskinesia in a quadriplegic distribution. It is of note also, that
children with spastic hemiplegia will usually be in levels I and
II, children with spastic diplegia will be in levels II, III and IV
and those with quadriplegia will be in levels III, IV and V.
This study highlights some of the problems with the commonly used motor type and topography classifications. In collecting data for the VCPR, where multiple sources of information were accessed for any one child, discrepancies were
frequently encountered in the identification of the most prominent motor type. Indeed, when the cases were reviewed for this
study, common misclassifications were found based on the definitions presented in Table 2. It became clear from comparing
the VCPR with gait laboratory records that the five children
(1.5%) registered as having dyskinesia all had pure athetoid CP,
but a significant proportion of those classified as having a spastic motor type had some element of dyskinesia. For example,
many children considered to have spastic hemiplegia, were later
found to have a spastic pattern in the lower limb and a dyskinetic
pattern in the upper limb, when they presented for upper limb
interventions. Many adolescents with spastic quadriplegia were
found to have dyskinesia in the upper limbs. These findings are
of more than taxonomic importance. Dyskinesia may respond
unpredictably to surgery designed for spasticity but favourably
to medication.5–7
In addition to both random and systematic classification errors, the method used to ascertain cases has the potential to
seriously bias the results, particularly if some groups are underascertained relative to others. The birth prevalence rate of CP for
this cohort is 1.98/1000 live births and this is comparable to the
rate of 2.08/1000 determined through a pooling of European CP
registers.14 Despite an overall high level of case ascertainment, it
is still possible that children functioning at GMFCS level I may
be under-ascertained by the VCPR, as these children are the least
likely to access specialist services. In contrast, children classified as GMFCS level IV or V have significant functional disability and associated comorbidities that necessitate early clinical
assessment, and practically ensure their identification through
the VCPR. Despite being included on the VCPR, children who
Cerebral palsy in Victoria
died before the age of 6 years were excluded from this study,
and this may have resulted in an underestimation of GMFCS
V. Children in GMFCS levels II, III and IV develop significant
gait deviations, which prompt evaluation by a developmental
paediatrician, orthopaedic surgeon or by referral to the gait laboratory. These children are more likely to be reliably ascertained
by the register.
The comparison with two other cohorts demonstrates some
of these classification and ascertainment issues. Classification
by topographical distribution showed the greatest variability between the three studies. The Canadian study had a higher incidence of quadriplegia, reflecting the referral-based nature of
their population cohort. The Swedish study12 classified topography quite differently, defining spastic quadriplegia as ‘massive
total motor disability, the upper limbs at least as affected as the
lower limbs and non-ambulant’.
The poor reliability of topographical classification is due to
the inconsistency of distinctions between ‘severe diplegia’ and
quadriplegia, and between asymmetrical hemi-syndromes and
bilateral CP. The use of body diagrams to describe the topographical distribution in CP, based on the presence or absence
of spasticity or limb involvement, appears promising but requires validation. Efforts have also been made by the European
group to standardize topographical classifications and motor
type.15
The Victorian CP register is already an invaluable research
tool but clinicians need further education about the need to better define, assess and report dyskinesia when classifying motor
type. We need to further examine the possibility that significant
changes in motor type are occurring between early childhood,
when children are classified and entered on the register, and
later in childhood when some of the critical decisions regarding spasticity management and gait correction surgery must be
made.6,16
In conclusion, the term ‘cerebral palsy’, despite all its shortcomings, is worth retaining, although it might be better to use the
term ‘cerebral palsies’ to describe a CP phenotype that encompasses the enormous variability in motor type, topography and
gross motor function.3 We think that the best way of classifying
children with CP is a combination of motor type, topography
and gross motor function, according to the GMFCS. Although
the motor type and topographical classifications lack reliability,
they are familiar, widely used and clinically significant. In the
meantime, the GMFCS is a simple, intuitive and reliable tool to
classify gross motor function in children with CP. It has modest correlations with measures of speech, hearing, vision and
cognition and strong correlations with certain musculoskeletal problems.14 These will be examined in future studies. The
GMFCS is recommended as a tool for the improvement of communication between health professionals involved in the care of
children with CP.
483
ACKNOWLEDGEMENTS
We acknowledge the contribution of Chris Blackburn, physiotherapist, from the Monash Medical Centre, and Mary Sheedy,
from the Department of Orthopaedics at the Royal Children’s
Hospital.
REFERENCES
1 Stanley F, Blair E, Alberman E. Cerebral Palsies: Epidemiology
and Causal Pathways. London: Mac Keith, 2000. Clinics in Developmental Medicine No. 151.
2 Mutch L, Alberman E, Hagberg B, Kodama K, Perat MV. Cerebral
palsy epidemiology: Where are we now and where are we going?
Dev. Med. Child Neurol. 1992; 34: 547–51.
3 Miller G, Clark GD. The Cerebral Palsies: Causes, Consequences,
and Management. Boston: Butterworth-Heinemann, 1998.
4 Dobson F, Boyd RN, Parrott J, Nattrass GR, Graham HK. Hip
surveillance in children with cerebral palsy. Impact on the surgical
management of spastic hip disease. J. Bone Joint Surg. Br. 2002;
84: 720–6.
5 Sanger TD, Delgado MR, Gaebler-Spira D, Hallett M, Mink JW.
Classification and definition of disorders causing hypertonia in
childhood. Pediatrics 2003; 111: e89–97.
6 Sanger TD. Pathophysiology of pediatric movement disorders. J.
Child Neurol. 2003; 18: S9–24.
7 Delgado M, Albright A. Movement disorders in children: Definitions, classifications and grading systems. J. Child Neurol. 2003;
18: S1–8.
8 Blair E, Stanley F. Interobserver agreement in the classification of
cerebral palsy. Dev. Med. Child Neurol. 1985; 27: 615–22.
9 Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi
B. Development and reliability of a system to classify gross motor
function in children with cerebral palsy. Dev. Med. Child Neurol.
1997; 39: 214–23.
10 Wood E, Rosenbaum P. The Gross Motor Function Classification
System for cerebral palsy: A study of reliability and stability over
time. Dev. Med. Child Neurol. 2000; 42: 292–6.
11 Gorter JW, Rosenbaum PL, Hanna SE et al. Limb distribution,
motor impairment, and functional classification of cerebral palsy.
Dev. Med. Child Neurol. 2004; 46: 461–7.
12 Nordmark E, Hagglund G, Lagergren J. Cerebral palsy in southern
Sweden II. Gross motor function and disabilities. Acta Paediatr.
2001; 90: 1277–82.
13 Bax M. Terminology and classification of cerebral palsy. Dev. Med.
Child Neurol. 1964; 6: 295–307.
14 Kennes J, Rosenbaum P, Hanna SE et al. Health status of schoolaged children with cerebral palsy: Information from a populationbased sample. Dev. Med. Child Neurol. 2002; 44: 240–7.
15 Surveillance of Cerebral Palsy in Europe. Prevalence and characteristics of children with cerebral palsy in Europe. Dev. Med. Child
Neurol. 2002; 44: 633–40.
16 Flett P. Rehabilitation of spasticity and problems in childhood cerebral palsy. J. Paediatr. Child Health 2003; 39: 6–14.